Publishing@W3C EPUB Survey Results

In February of this year, the Publishing@W3C community ran a survey for the publishing industry on the topic of EPUB and the future of publishing. The main focus of the survey was to understand how people are using EPUB, what challenges they face, or what they wish they could do but currently cannot. The other goal of the survey was to understand more about the needs of readers and members of industries like educational publishing.

The survey in total was 85 questions and we received over 250 responses! The Publishing@W3C community would like to thank everyone who responded, with this data we can be much more decisive in determining what the future of publishing looks like.

The full results of the survey can be found here, in this blog post I will share some highlights and interesting findings.

Reading Systems and Readers

Some of the questions in the survey focused on the needs and problems faced by reading system developers and users. We learned that only 5 of 14 respondents have implemented all of the EPUB specifications, with the majority citing a lack of ROI (return on investment) being the reason not to implement outside of the core specification. Another interesting comment was that since many features exist as part of WebAPIs for browser-based renderers, there’s no return on building EPUB-specific functionality.

The most difficult parts of the specification to implement include Media Overlays, Fixed Layout, and Javascript.

The good news is that reading systems are eager for more: 83% reported they’d like to have more conformance information in the specification, and an overwhelming 95% would love to see tests. They’re also interested in accessibility, while only 37% reported they display provided accessibility metadata, 53% of those who do not currently say they plan on doing so in the future.

There are still challenges out there, reading systems are reported to be difficult to do accessibility testing with, interoperability is a huge challenge for publishers and users, as one respondent put it: “EPUB is fine, but reading systems limit it’s potential”. Content creators can’t leverage modern web stack features because of inconsistent/nonexistent support.

EPUB Today

One of the largest sections of the survey focused on how ebook developers and publishers use EPUB. The EPUB specification is quite feature-rich, not counting the satellite specifications, and we wanted to get an idea of exactly how many of those features are used. We also wanted to know how many people made the switch to EPUB3.

72% of those who responded publish in EPUB3, with an additional 20% publishing a combination of EPUB2 and EPUB3. Just over half support a backlist of EPUB2 content as well. Just under half of those who responded still produce PDF content on top of their EPUB content, citing partner requirements and formatting as the main reasons for supporting the multiple formats.

Fixed layout is still the content type of choice for children’s books, graphic novels, and some educational content. When asked why more people do not create high-design reflowable publications instead, respondents cited the need for support of modern CSS layouts like Grid or Flexbox, support for full-bleed images in reflowable books, and cost of production.

A majority of respondents believe EPUB needs more core media types, in keeping with the modern web platform. 74% cited an interest in using scripting (Javascript) in EPUB if it were better supported by reading systems and the specification. Features like user interactivity in quizzes, charts, images and maps were the most popular. MathML is used by a small group of content creators, but many mentioned the need for better support from reading systems.

A big question we wanted to answer was who was using epub:type and how. 79% reported that they do use epub:type, with the most common uses being for document partitions, navigation, divisions, sections, and annotations. The main expectation users had from employing `epub:type` was that assistive technologies would be able to use the types to process the semantic information. The other expectations were for annotations appearance and navigation for reading systems.

Overall, there were no huge surprises in the EPUB ecosystem, with potentially the exception of the usage of epub:type. EPUB3 is a strong format, and while there are many ways we can improve the experience of users and content creators, the foundation we’re working from is solid.

EPUB Tomorrow

As mentioned before, there are still ways to improve EPUB, and we asked the community about what they would like to see in future versions or releases of EPUB. Some of these ideas will be taken to the proposed EPUB3 Working Group and others will be sent to the Publishing Community Group for further discussion and incubation.

When looking at the future of EPUB, we wanted to identify what things people struggled or were completely unable to publish using the format. EPUB is flexible, and many of us have seen it used in amazing ways, but it does not work for every use case, and to understand where we could improve, we needed to know what those use cases were.

When asked what features content creators would like to see supported in EPUB, the top results were Dual Language Support, notifications for readers, background or full bleed images, and more interactive features like testing or map integrations. 63% of respondents expressed an interest in mixing fixed layout and reflowable sections in an EPUB. For those who are keen on the specification, you’ll know this is technically already possible, but has very little implementation.

The last big request for improvement was around data formats and scripting. EPUB3 has long supported javascript as a scripting format, but implementation is mixed at best and the specification does not address it beyond explaining how to reference it in the content documents.

Many things asked for in the survey are features EPUB already supports or theoretically could, but aren’t laid out explicitly in the specification. One of the goals of the new EPUB 3 Working Group is to address those gaps, be they issues with implementations or a lack of clarity in the language of the spec. We look forward to addressing as many of these problems as we can. Of course, if there is anything we’re missing, please feel free to let us know by filing an issue in the GitHub repository.

We want to thank all of the respondents again for their feedback and time, it has been immensely helpful for gaining an understanding of our industry today. We will be in touch with anyone who expressed an interest in doing so, and for anyone who did not have a chance to complete the survey, or would like to know more about it or anything related to Publishing@W3C, please feel free to reach out to us on twitter @w3cpublishing or by email at [email protected].